U.S. patent application number 11/754140 was filed with the patent office on 2008-01-03 for methods for providing an easily comprehendible risk rating for pharmaceutical products.
Invention is credited to Hugo Stephenson.
Application Number | 20080005174 11/754140 |
Document ID | / |
Family ID | 38878015 |
Filed Date | 2008-01-03 |
United States Patent
Application |
20080005174 |
Kind Code |
A1 |
Stephenson; Hugo |
January 3, 2008 |
Methods for Providing an Easily Comprehendible Risk Rating for
Pharmaceutical Products
Abstract
The invention relates to a method for assigning a risk rating to
a medical product. The method includes assessing one or more
threats associated with the medical product; assessing the level of
experience with the medical product; and assigning a risk rating
for the medical product to provide an indication of risk associated
with the medical product. The method may be implemented as a
webpage. For example, a new contraceptive may be assigned a risk
rating of yellow for the general treatment population. This risk
rating allows consumers to make an informed choice between
different products on the basis of benefit versus risk, and help
patients decide what steps they may wish to take to minimize their
risk if they choose to take the new drug.
Inventors: |
Stephenson; Hugo;
(Princeton, NJ) |
Correspondence
Address: |
WILLIAM DOUGLAS HARE
3 ANDERSON LANE
PRINCETON
NJ
08540
US
|
Family ID: |
38878015 |
Appl. No.: |
11/754140 |
Filed: |
May 25, 2007 |
Current U.S.
Class: |
1/1 ; 705/1.1;
707/999.107 |
Current CPC
Class: |
G06Q 50/00 20130101;
G16H 10/20 20180101 |
Class at
Publication: |
707/104.1 ;
705/1 |
International
Class: |
G06Q 50/00 20060101
G06Q050/00; G06F 17/40 20060101 G06F017/40 |
Claims
1. A method for assigning a risk rating to a medical product, the
method comprising: assessing one or more threats associated with
the medical product; assessing the level of experience with the
medical product; and assigning a risk rating for the medical
product to provide an indication of risk associated with the
medical product.
2. The method of claim 1, wherein the medical product comprises one
or more of a pharmaceutical product, a biologic product, and a
medical device.
3. The method of claim 1, wherein assessing one or more threats
comprises assessing the severity of the threat.
4. The method of claim 3, wherein assessing the severity of the
threat comprises one or more of a risk of permanent disability,
death, and serious adverse event.
5. The method of claim 1, wherein assessing one or more threats
comprises assessing the probability of the threat occurring.
6. The method of claim 1, wherein assessing one or more threats
comprises assessing the potential population affected by the
threat.
7. The method of claim 6, wherein assessing the potential
population affected by the threat comprises assessing a percentage
of the population to which the threat applies.
8. The method of claim 6, wherein assessing the one or more threats
comprises one or more of determining the population to which the
threat applies, determining whether the population at risk as a
result of the threat can be identified in advance, and determining
an indicator of usage of a second medical product for which there
is an interaction with the first medical product.
9. The method of claim 1, further comprising pooling the assessment
of each threat.
10. The method of claim 1, further comprising assigning an
indication of a level of usage of the medical product.
11. The method of claim 1, further comprising determining whether
the medical product is an orphan product.
12. The method of claim 1, wherein the risk rating comprises a
gradated rating having more than two grades.
13. The method of claim 12, wherein the risk rating comprises at
least three grades.
14. The method of claim 13, wherein the three grades comprise three
symbols.
15. The method of claim 1, wherein the method assesses more than
one threat.
16. The method of claim 1, wherein assigning a risk rating further
comprises providing guidance associated with the risk rating.
17. The method of claim 1, wherein the method for assigning a risk
rating to a medical product is implemented on a webpage.
18. A webpage for providing a risk rating for a medical product,
the webpage configured to: receive input relating to a drug; and
display a risk rating for the drug, the webpage including
instructions for assessing one or more threats associated with the
medical product, assessing the level of experience with the medical
product, and assigning a risk rating for the medical product to
provide an indication of risk associated with the medical
product.
19. The webpage of claim 18, wherein the medical product comprises
one or more of a pharmaceutical product, a biologic product, and a
medical device.
20. The webpage of claim 18, wherein assessing one or more threats
comprises assessing the severity of the threat.
21. The webpage of claim 20, wherein assessing the severity of the
threat comprises one or more of a risk of permanent disability,
death, and serious adverse event.
22. The webpage of claim 18, wherein assessing one or more threats
comprises assessing the probability of the threat occurring.
23. The webpage of claim 18, wherein assessing one or more threats
comprises assessing the potential population affected by the
threat.
24. The webpage of claim 18, wherein the risk rating comprises a
gradated rating having more than two grades.
25. The webpage of claim 24, wherein the risk rating comprises at
least three grades.
26. The webpage of claim 25, wherein the three grades comprise
three symbols.
27. The webpage of claim 18, wherein the webpage assesses more than
one threat.
28. The webpage of claim 18, wherein assigning a risk rating
further comprises providing guidance associated with the risk
rating.
Description
TECHNICAL FIELD
[0001] The field of the invention generally relates to methods,
software, systems, and webpages for providing consumers,
physicians, regulators, and manufacturers with a risk rating for
pharmaceutical products, biologics and medical devices.
BACKGROUND
[0002] Drugs are approved by the US Food and Drug Administration
(FDA) after undergoing clinical studies to show safety and
effectiveness. The clinical studies involve phase I studies on a
small population, approximately thirty patients, to demonstrate
safety. Next the drug is put through phase II studies to determine
dosing levels. These studies are on a larger population of
approximately 100 to 300 patients. With the dosing levels
established, the drug is studied on a significantly larger
population of 3000 to 5000 patients to show effectiveness and look
for indications that there may be safety concerns with the drug.
After the drug is approved, the FDA may require post-approval, or
phase IV, studies to obtain a more complete safety profile of the
drug.
[0003] The safety and effectiveness information gathered from the
clinical studies conducted to obtain approval are used to create
the label. This information in the label, however, is not readily
comprehendible by most patients, may be overwhelming, and is
unlikely to inform the patient as to how the label applies to them
as an individual. For example, the adverse events that a person
would read in a typical drug label range from death to dizziness.
Such a broad range of adverse events does not adequately inform a
person of which adverse events apply to them.
[0004] Some US FDA-approved labels have a type of warning known in
the industry as a black box warning. Examples of such drugs with
black box warnings are Accutane.RTM., the various generic versions
of isotretinoin, and thalidomide. The black box warning is
prominently displayed at the front of the FDA-approved labeling and
is in the form of a black box surrounding text. In the
Accutane.RTM. black box warning, the text describes precautions
that should be taken when prescribed the drug, warnings about the
drug and its potential adverse effects, special prescribing
information and contraindications for the drug. The Accutane and
isotretinoin black box warnings are directed to the threat of birth
defects if women take the drug while pregnant. It should be noted
that other threats are alleged to be associated with isotretinoin,
including suicide, but only the threat of birth defects is included
in the black box warning. It should be recognized that the black
box warning is at most a binary indicator of a threat associated
with the drug: its presence indicates the existence of the warning
and its absence indicates only that a black box warning has not yet
been required for the drug.
[0005] While a label informs the patient about the information
known about the drug at one particular time, the label does not
inform the patient about what is not yet known about the drug. It
is not uncommon for the safety profile of pharmaceuticals to more
completely emerge only after millions of patients have used the
drug. The information that contributes to that safety profile
typically emerges in a piece-meal fashion with sporadic adverse
events reported and occasional studies published. For example, two
drugs with similar labels for the same indication may nonetheless
have different levels of risk that cannot be determined by a
patient from reading the label because one drug has been used by
millions more patients than the other drug and therefore its safety
profile is more clearly known. Addressing in part the different
levels of risk associated with newly approved drug, in the United
Kingdom newly approved drugs have a symbol on their label to
indicate that it is a newly approved drug.
[0006] Information that relates to a pharmaceutical's safety
profile may be gathered from drugs in that class rather than the
drug itself. For example, safety information relating to
pharmaceuticals such as Baycol, a statin, and Vioxx, a COXII
inhibitor, had implications for other pharmaceuticals in both
classes. With respect to Vioxx, this information was not generally
known to the patient community as the information developed, but
only upon an announcement by the company after a large quantity of
data had been gathered. The inventor has developed a system to keep
consumers of pharmaceuticals more fully informed in a real-time
manner about the risk associated with the pharmaceuticals that they
are prescribed. Advantageously, the inventor has developed systems
to provide a risk rating in a manner that is easily
comprehendible.
[0007] Similar concerns exist for biologics and medical devices.
For example, drug coated stents have been the subject of safety
questions as a result of clinical reports. In the past the Dalcon
Shield and silicone breast implants were the subject of safety
concerns. Thus, there exists a similar need to inform consumers in
a readily comprehendible manner about risk relating to the medical
devices and biologics they may be using.
SUMMARY
[0008] In one general aspect, a method for assigning a risk rating
to a medical product, the method includes:
[0009] assessing one or more threats associated with the medical
product;
[0010] assessing the level of experience with the medical product;
and
[0011] assigning a risk rating for the medical product to provide
an indication of risk associated with the medical product.
[0012] Embodiments of the method may include one or more of the
following features. For example, the medical product may include
one or more of a pharmaceutical product, a biologic product, and a
medical device.
[0013] Assessing one or more threats may include assessing the
severity of the threat. Assessing the severity of the threat may
include one or more of a risk of permanent disability, death, and
serious adverse event. Assessing one or more threats may include
assessing the probability of the threat occurring.
[0014] Assessing one or more threats may include assessing the
potential population affected by the threat. Assessing the
potential population affected by the threat may include assessing a
percentage of the population to which the threat applies.
[0015] Assessing the one or more threats comprises one or more of
determining the population to which the threat applies, determining
whether the population at risk as a result of the threat can be
identified in advance, and determining an indicator of usage of a
second medical product for which there is an interaction with the
first medical product.
[0016] The method may further include pooling the assessment of
each threat. The method may further include assigning an indication
of a level of experience with the medical product. The method may
further include determining whether the medical product is an
orphan product.
[0017] The risk rating may be implemented as a gradated rating
having more than two grades. The risk rating may be implemented
with at least three grades. The three grades may be in the form of
three symbols.
[0018] The method may assess more than one threat. Assigning a risk
rating may further include providing guidance associated with the
risk rating. The method for assigning a risk rating to a medical
product may be implemented on a webpage.
[0019] In another general aspect, there is provided a webpage for
providing a risk rating for a medical product. The webpage may be
configured to receive input relating to a drug; and display a risk
rating for the drug. The webpage includes instructions for
assessing one or more threats associated with the medical product,
assessing the level of experience with the medical product, and
assigning a risk rating for the medical product to provide an
indication of risk associated with the medical product.
[0020] Embodiments of the webpage may include one or more of the
following features. For example, the medical product may be one or
more of a pharmaceutical product, a biologic product, and a medical
device.
[0021] Assessing the one or more threats may include assessing the
severity of the threat. Assessing the severity of the threat may
include one or more of a risk of permanent disability, death, and
serious adverse event.
[0022] Assessing one or more threats may include assessing the
probability of the threat occurring. Assessing one or more threats
may include assessing the potential population affected by the
threat.
[0023] The risk rating may be implemented as a gradated rating
having more than two grades. The risk rating may be implemented
with at least three grades. The three grades may be in the form of
three symbols. The webpage may assess more than one threat.
Assigning a risk rating may further include providing guidance
associated with the risk rating.
[0024] The details of various embodiments of the invention are set
forth in the accompanying drawings and the description below. Other
features and advantages of the invention will be apparent from the
description, the drawings, and the claims.
DESCRIPTION OF THE DRAWINGS
[0025] FIG. 1 is a flowchart for an algorithm to establish a risk
rating for a drug based on the threats associated with the drug and
the level of experience with the drug.
[0026] FIG. 2 is an implementation of step 15 of the flowchart of
FIG. 1 for rating the individual threats known to be associated
with a drug.
[0027] FIG. 3 is an implementation of the flowchart of FIG. 2 for
an algorithm for assigning penetration ratings for individual
threats.
[0028] FIG. 4 is an implementation of the flowchart of FIG. 2 for
an algorithm for assigning threat levels for individual
threats.
[0029] FIG. 5 is an implementation of the flow chart of FIG. 1 for
an algorithm for assigning a risk rating for a drug given
previously assigned threat levels and penetration ratings for
different known threats, and the level of experience that exists
for the drug.
DETAILED DESCRIPTION
[0030] The inventor has developed systems for providing real-time
risk ratings, or safety profile indicators, about medical products,
such as pharmaceuticals. The systems provide a risk rating for the
pharmaceutical in a manner that is easily comprehendible to
consumers, physicians, regulators, and manufacturers. In one
aspect, the risk rating is gradated to distinguish between
different levels of risk associated with a drug. This is in
contrast to the black box warning described above which is not
gradated but instead is binary: present or absent. In another
aspect, the risk rating is based on pooling multiple threats rather
than a single threat. Again, this is in contrast to the black box
warning described above which considers a single threat--birth
defects associated with taking the drug while pregnant.
[0031] Although equally applicable to medical devices and
biologics, for ease of explanation the system is described only
with respect to pharmaceuticals. As described in greater detail
below, on a most basic level, the system uses an algorithm to rate
individual, known threats to safety associated with a drug, pool
these individual threat levels, assess the level of experience that
exists for the drug, and assign a risk rating for the drug. In this
process, the systems assess a drug based on what is known and not
known about the drug, and provide a meaningful categorization of
risk. The system can be applied to the general population to assign
a general risk rating, or to a sub-population in order to assign a
personalized risk rating for an individual.
[0032] As a benefit to patients, the risk ratings are configured to
be easily comprehendible and may be visually-based (e.g., letter,
word, color, number(s), shape, etc.). The systems optionally
provide guidance for patients and physicians regarding additional
actions that may need to be taken. The ratings are determined as a
function of at least one or more of the following factors: the
probability of different threats occurring, their potential
severity, the proportion of people potentially affected by the
threats, and the level of experience available for the drug. Other
factors may be included in the function as needed to improve any
aspect of the risk rating.
[0033] The following terms used in the application are defined as
follows. A threat is an individual item of safety information that
exists for a drug and can be derived from one or more sources,
e.g., the labeling, a clinical study reported in a journal article,
an FDA alert on the drug or even class of drugs of which the
subject drug is a member, a news article, an anecdotal report in a
blog, etc. Moreover, there may be multiple threats contained within
a single source of information. For example, a particular drug's
label may have warnings, precautions, contraindications, adverse
reactions, drug-drug interactions, and other threats. Within the
warnings in the labeling there may be separate warnings relating to
a number of factors, such as liver function, affect on blood
pressure, etc. Each of these individual factors is considered a
threat. Similarly, if a journal article reports on a large scale
post-marketing clinical study of a drug, the article may report
that certain subpopulations were the subject of a particular safety
concern. Each of these individual safety concerns is considered a
threat. Other subpopulations within the study may have had
drug-drug interactions. Again, each of these drug-drug interactions
is considered a threat.
[0034] FIG. 1 is a flowchart for an algorithm 10 to calculate a
risk rating for a drug. The algorithm 10 includes an assessment of
individual known threats associated with a drug to rate each
individual threat (step 15). In assessing the individual threats
associated with a drug, the algorithm takes into account factors
such as (a) absolute and/or relative size of the population that
will be affected by the threat; (b) severity of the outcome if the
threat should occur; and (c) probability of the threat occurring.
These are but some factors that can be included in assessing the
threats and not all of the factors must be used.
[0035] The assessment of the factors can be performed using
different weightings. In particular, the weighting of the factors
can be adjusted on a subjective basis depending upon the emphasis
intended in setting up the algorithm 10. For example, in one
possible implementation, more weight can be put on the severity of
the outcome than on the size of the population or the probability
of the threat occurring.
[0036] After rating the individual, known threats (step 15), the
algorithm then pools the ratings of the threats existing for the
drug (step 20). The pooling can be configured in a number of
manners. In one basic implementation, if the ratings are based on a
numerical ordering the pooling can be configured to sum the
assigned values and divide by the number of values. As one example,
the algorithm can be used to evaluate only the probability of an
occurrence. In that algorithm a low probability may be assigned a
rating value of one, a medium risk may be assigned a rating value
of two and a high risk may be assigned a rating value of three. In
that scenario there are a total of twelve threats broken down into
three threats of value one, four threats of value two and five
threats of value three, and the pooling would give a rating value
of 2.167 for the drug, which is between a medium and a high
probability. In this implementation, the pooling is based on the
assumption that a probability level for a drug is based on an
average of the probability levels determined for the individual
threats. In another implementation using the same rating values for
the individual probability levels, the output may be configured to
be more cautious. In such an implementation, the output may be
determined based on the assumption that the probability level
should be set at the highest level that includes at least a certain
percentage, e.g., 25%, of the probability levels determined for
individual threats. Using the above data, where five of twelve
(i.e., 42%) of the probabilities are high probability, the
probability level pooling gives a rating value of high probability
for the drug. If a total of three out of twelve, (i.e., 25%) of the
individual probability levels determined were a combination of
medium and high level, the probability level pooling would give a
medium level of probability for the drug.
[0037] In yet another implementation, based on extreme caution, the
pooling would be configured to give a rating corresponding to the
highest level of probability that any one threat exhibits. Thus if
eleven of the twelve are low probability but one of the twelve is a
high probability, in this model of extreme caution the pooling
would give a rating value corresponding to a high level of
probability to the drug.
[0038] It should be understood that level of probability of a
threat occurring is but one factor that could be used in rating the
threats and other factors could be used in the pooling step. For
example, in yet another implementation, a severity rating used with
the individual, known threat may be in the form of a letter, e.g.,
D for the highest level of severity, C with a lesser level of
severity, B with an even less level of severity, and A with the
least level of severity. In pooling the individual threats, the
letter rating for severity could be used with another output from
the pooling step. Thus, if the probability level pooled is a
percentage (i.e., ranging between 0% to 100%), a particular pooling
may give a value of D75 to indicate a threat with the highest level
of severity and a 75% probability.
[0039] The algorithm 10 also can use an assessment of the level of
experience with a drug (step 25) in calculating a risk rating for
the drug (step 30). Experience can be measured in a number of
manners, such as years on the market, number of prescriptions,
number of patients studied during the clinical trials, number of
patients taking the drug, number of countries in which the drug has
been approved and marketed, etc. These factors relating to
experience are designed to capture the safety knowledge that is
gained through use of the drug as well as have an effect on the
assessment if there is relatively little experience with the
drug.
[0040] The algorithm 10 then assigns a risk rating for the drug
(step 30) based on a combination of the pooling of threat ratings
for the drug (step 20) and assessment of the level of experience
with the drug (step 25). In one implementation of the risk rating
step, the resulting risk rating of the algorithm 10 is defined
according to the system illustrated in Table 1.
TABLE-US-00001 TABLE 1 Risk Rating Risk Rating Risk Level Patient
Advisory RED HIGH Should only be used under strict supervision
ORANGE ELEVATED Use within a customized risk management plan YELLOW
GUARDED Be on the lookout for safety events BLUE GENERAL Use under
the normal care of a physician GREEN LOW Suitable for widespread
use
[0041] FIG. 2 provides a more detailed explanation of the rating
step 15 of FIG. 1, implemented as an algorithm 50. Initially the
algorithm 50 is used to evaluate the severity of a threat (step
55). Factors that may be used to evaluate severity of the threat
include the outcome expected if the threat occurs. The expected
outcome may be found in the labeling, from the source of the threat
(e.g., described in the journal article that reported the threat),
or based on a medical judgment. Examples of outcomes include
permanent disability, death, and serious adverse effect. Other
factors used in evaluating severity include whether the risk level
of the outcome is high, elevated, or a general warning.
Consequently, the result of the evaluation of step 55 will be
either a description of the outcome, e.g., permanent disability, or
an indicator of the outcome, e.g., a numerical value of 1 for
death, a numerical value of 2 for permanent disability, etc.
[0042] The algorithm 50 also evaluates the probability of the
threat occurring (step 60). This evaluation can be a specific
probability (1 in 10,000) or a probability range (between 1 in
10,000 and 1 in 20,000, etc.). The probability value can be based
on a number of methods, such as reviewing historical data (e.g.,
the clinical studies used to obtain FDA approval), epidemiological
studies, or a medical judgment. The result of step 60 can be in the
form of a probability or an indicator of probability. One example
of an indicator of probability is the use of terms such as high,
elevated, and low.
[0043] Next, the algorithm 50 assigns a threat level (step 65) that
is based on both the severity of the threat (step 55) and the
probability of the threat (step 60). The threat level can be based
on equal weights for the severity of the threat and the probability
of the threat occurring. Such an equal weighting, however, is not
necessary and can be in a range that varies from being based
entirely on one of the two factors to the other of the two factors.
Moreover, the threat level can be based on other, different
weightings of each factor. This assignment of a threat level is
described in more detail below.
[0044] Along with assigning a threat level (step 65), the algorithm
also assigns a penetration rating based on the size of the
population affected by the threat (step 70). This assignment of a
penetration rating is described in more detail below.
[0045] The algorithm 50 next assigns a threat rating for each
particular threat being evaluated (step 75). The individual threat
rating calculated or otherwise determined will be a combination of
the inputs from steps 65 and 70. The threat rating assigned may be
in the form of an indicator, the form of the indicator being
varied. For example, the indicator may be a color, shape, number,
text, code, etc. For example, in one implementation, the indicator
may merely be the merging of the penetration rating (step 70) and
the threat level (step 65). Thus, if the penetration rating is 85%
and the threat level is a letter C, the indicator resulting from
step 75 may be in the form of an 85 C. Alternatively, the
penetration and risk levels can be assigned numbers that can be
combined in some manner to imply the individual threat rating.
Thus, if the maximum penetration rating that can occur is given a
value of 50 and the highest threat level that can be assigned is a
value of 50, then the highest individual threat rating possible is
a 100 if the two values are added, or a 50 if the two values are
averaged.
[0046] The penetration rating of step 70 of the algorithm 50
illustrated in FIG. 2 can be implemented using a variety of factors
to assign a penetration rating for a particular threat. For
example, in one implementation of this portion of the algorithm,
the penetration rating may be assigned as illustrated in Table II,
below.
TABLE-US-00002 TABLE II Categorization of Level of Penetration of
Population Affected by a Threat Penetration Rating Factors 85 If
the threat applies to a sub-population > 85% of the treatment
population If population at risk cannot be identified in advance If
the threat involves an interacting drug with an OTC drug 15 If the
threat applies to a sub-population > 15% treatment population If
the threat involves an interacting drug that is a high volume
prescription drug in the treatment population 1 If the threat
applies to a sub-population > 1% treatment population If the
threat involves an interacting drug that is a non-high volume,
non-specialist drug in the treatment population 0 If the threat
applies to a sub-population < 1% treatment population If the
threat involves an interacting drug that is a specialist only drug
in the treatment population
[0047] These are exemplary of the categorizations of the
penetration and the factors that can be used to determine the
population potentially or actually affected. The factors in Table 2
relate to three factors: the amount of a sub-population to which
the threat applies, whether the population at risk can be
identified in advance, and the nature of an interacting drug if the
threat involves a drug-drug interaction. As can be expected, these
factors and categorizations can be modified and/or replaced
depending upon the design of the assessment and algorithm.
[0048] In the flowchart illustrated in FIG. 3, a penetration
algorithm 100 determines the extent to which a threat applies to a
population of patients. Applying the penetration algorithm 100
illustrated in FIG. 3 to a particular threat results in the threat
being categorized as having one of four penetration ratings: 85,
15, 1 or 0. It should be noted that the penetration of a particular
threat within a specific population can be rated in numerous
manners. The penetration can be divided, for example into four
groups: MOST, MANY, SOME, or FEW; or divided into two groups: most
and few, or more than half and less than half. The penetration
rating also can be made up of narrower ranges or greater divisions:
less than one third of the patients, greater than one third of the
patients but less than two thirds of the patients, and greater than
one third of the patients but less than all of the patients. The
intent is to apply the algorithm 100, at least in part, based on
the number of patients potentially or actually subject to the
threat.
[0049] Initially, the penetration rating algorithm 100 examines
whether the threat applies to a particular proportion of the
treatment portion (step 105). In the implementation illustrated in
FIG. 3, the algorithm is set to determine whether the threat
applies to a sub-population that is greater than 85% of the
treatment population. If the threat applies to a sub-population
that is greater than 85% of the treatment population, then the
algorithm assigns a penetration rating to the threat of 85 (step
120). If the condition is not true, the algorithm determines
whether the population at risk can be identified in advance. If the
population cannot be identified in advance, then the algorithm
assigns a penetration rating to the threat of 85 (step 120). One
rationale for equating the same penetration rating for 85% of the
population and not being able to identify the population in advance
is that one should set the penetration rating at a high level when
the population at risk cannot be identified in advance.
[0050] If the population can be identified in advance, the
algorithm determines whether the threat involves an interacting
drug that is an over-the-counter (OTC) drug (step 115). An OTC drug
can be purchased without a doctor's prescription and thus without a
doctor or pharmacist's intervention to warn the patient about
potential interactions. Thus, the penetration rating is given the
highest penetration rating possible, namely, 85 (step 120) in this
algorithm. Of course, any other indicator of penetration may be
used, such as the absolute number of prescriptions written for the
interacting drug, prescriptions written on an annual basis for the
interacting drug, the number of prescriptions written for the
entire class of drugs in which the interacting drug is a member, or
a basis that provides a means to differentiate between levels of
prescriptions written for a drug or class of drugs.
[0051] It should be noted that steps 105, 110, and 115 are arranged
in one particular order. This arrangement, however, is not to be
construed as limiting the method to this particular order. In fact,
the order of these three steps can be reversed or re-ordered and
give the same result when applied to the same set of
conditions.
[0052] If the threat does not involve a factor that causes the
algorithm to assign a penetration rating of 85, the algorithm 100
next determines whether the threat should be assigned a penetration
rating of 15 (steps 125, 130, 135). The algorithm determines
whether the threat applies to a sub-population that is greater than
fifteen percent of the treatment population (step 125) or involves
an interacting drug that is a high volume prescription drug (step
130). If either of these conditions is true, the algorithm assigns
a penetration rating of 15 (step 135). While a high volume
prescription drug is highly prescribed, it nonetheless is subject
to a doctor or pharmacist's review before being taken by a patient.
Thus, a physician and/or pharmacist will be aware of, or can
inquire about, the other drugs being taken by the patient. This
will reduce the population that takes the interacting drug along
with the drug that is the subject of the threat and for which the
algorithm 100 is being processed. Although steps 125 and 130 result
in the same penetration ratings, the algorithm 100 can be
configured to assign different penetration ratings to each
step.
[0053] If the threat does not involve a factor that causes the
algorithm to assign a penetration rating of 15 in steps 125 or 130,
the algorithm 100 next determines whether the threat should be
assigned a penetration rating of 1 (steps 140, 145, 150). The
algorithm determines whether the threat applies to a sub-population
that is greater than one percent of the treatment population (step
140) or involves an interacting drug that is a non-high volume
prescription drug (step 145). If either of these conditions is
true, the algorithm assigns a penetration rating of 1 (step 150).
Similar to steps 125 and 130, the algorithm 100 can be modified to
assign different penetration rating to steps 140 and 145.
[0054] If the threat does not involve a factor that causes the
algorithm to assign a penetration rating of 1, the algorithm 100
next determines whether the threat should be assigned a penetration
rating of 0 (steps 155, 160, 165). The algorithm determines whether
the alert applies to a sub-population that is less than one percent
of the treatment population (step 155) or involves an interacting
drug that only a specialist will prescribe (step 160). If either of
these conditions is true, the algorithm assigns a penetration
rating of 1 (step 165).
[0055] Referring to FIG. 4, an algorithm 200 implements step 65 of
the algorithm 50 (FIG. 2) to calculate a threat level as an
indicator of risk. In the implementation illustrated in FIG. 4, the
threat level assigned is based on the letters A, B, C and D,
although other indicators of risk, such as a color coded system,
can be used instead. Table III is a table showing the
implementation used in FIG. 4 to assess the risk of the threat and
assign an indicator of risk for each threat. Table III also
provides examples of the threats that correspond to the risk level
and indicator of risk. For the purposes of Table III, the terms
HIGH RISK and ELEVATED RISK are defined as follows: (a) HIGH RISK:
a likelihood of permanent disability (PD) or death of 1/10,000 or
other serious adverse effect (SAE) of 1/1,000; and (b) ELEVATED
RISK: a likelihood of PD or death of 1/100,000 or other SAE of
1/10,000. The output of the safety algorithm is the assignment of
threat level.
TABLE-US-00003 TABLE III Threat Level Indicator Threat Risk Level
Level Examples of Threat HIGH risk of PD/death D Absolute
contraindications, CLASS 1 (ORCA) DDIs, black box warnings HIGH or
ELEVATED risk of C e.g. insulin, erythropoetin, PD/death if drug
dosing/titration warfarin not closely monitored HIGH risk of SAEs C
Relative contraindications, warnings ELEVATED risk of PD/death C
Precautions ELEVATED risk of SAEs B Precautions General warning A
General label change
[0056] To implement the safety algorithm 200, the algorithm
initially determines whether the threat implies a high risk of
permanent disability ("PD") or death (step 215). As noted above, a
high risk of permanent disability or death is defined in terms of
probability of the event associated with the threat occurring or
any other serious adverse effect occurring. If the threat implies a
high risk of permanent disability or death, the algorithm assigns a
threat level of D (step 220).
[0057] If the threat does not imply a high risk of permanent
disability of death, the algorithm 200 will determine whether the
threat involves a HIGH or ELEVATED risk of permanent disability or
death if the drug dosing or titration is not closely monitored
(step 225). Examples of drugs for which this condition may be true
include insulin, erythropoetin, and warfarin. If the condition is
true, the algorithm assigns a threat level of C (step 240). If the
condition is not true, the algorithm determines whether the threat
implies a HIGH risk of serious adverse events occurring (step 230).
Again, the definition for HIGH risk described above is used in
determining the level of risk. If the condition is true, the
algorithm assigns a threat level of C (step 240). If the condition
is not true, the algorithm 200 determines whether the threat
implies an ELEVATED risk of permanent disability or death (step
235). The definitions of risk described above are used to determine
the level of risk, i.e., HIGH or ELEVATED, involved with the
threat. If the condition is true, the algorithm assigns a threat
level of C (step 240).
[0058] If the condition is not true, the algorithm 200 determines
whether the threat implies an ELEVATED risk of serious adverse
events (step 250). If the condition is true, the algorithm assigns
a threat level of B (step 245). If the condition is not true, the
algorithm 200 assigns a threat level of A (step 260). Thus, after
the application of the algorithm 200 to a particular threat, the
algorithm has assigned a threat level of either A, B, C or D for
that threat.
[0059] Referring to FIG. 5, a risk rating assignment algorithm 300
assigns a risk rating to indicate the level of risk for a
pharmaceutical based on the output of the algorithms 100 and 200
illustrated in FIGS. 3 and 4, respectively, and the level of
experience available for the drug. The algorithm 300 loosely
corresponds to step 30 of FIG. 1, in which the algorithm 10 assigns
a risk rating for a drug. The algorithm 300 is arranged to sort the
risk associated with the drug in decreasing levels of risk.
Referring to the conditional questions of step 305, the algorithm
determines whether a threat with a high threat level and
penetration rating exists for the pharmaceutical. In step 305 of
the algorithm, if a threat with a D threat level risk and
penetration rating of 85 exists, (from FIGS. 3 and 4, algorithms
100 and 200, respectively), the algorithm 300 assigns a red risk
advisory rating to the pharmaceutical overall. In FIGS. 3 and 4,
the penetration and threat level is assessed for each threat. In
FIG. 5, the algorithm 300 considers all of the assessments of the
individual threats determined in FIGS. 3 and 4 to assign a risk
rating to the drug overall.
[0060] If the condition of step 305 is not true, then the algorithm
300 determines the conditions set forth in step 310: does a threat
with a D threat level exist with a penetration rating of 1? As in
step 305, the input for assessing the condition of step 310 is the
output of the algorithms 100 and 200 of FIGS. 3 and 4,
respectively. If this condition is true, the algorithm assigns an
orange risk rating to the pharmaceutical. If the condition is not
true, the algorithm determines the condition of step 315: does a
threat with a C threat level exist with a penetration rating of 85
patients. If the condition of step 315 is true and a threat with a
C threat level exists with a penetration rating of 85, the
algorithm assigns an orange risk rating to the pharmaceutical.
[0061] If the condition of step 315 is not true, the algorithm 300
determines the condition of step 320: is the drug an orphan drug.
As illustrated in FIG. 5, if the drug is an orphan drug, the drug
is assigned an orange risk rating. One rationale for assigning an
orange risk rating to an orphan drug is the relatively small number
of people in whom the drug has tested and used.
[0062] If the drug is not an orphan drug (i.e., the condition of
step 320 is not true), the algorithm determines the condition of
step 325: is there less than two years of market experience with
the drug or have there been less than one million prescriptions
written for the drug. If the condition is true (i.e., less than two
years on the market or less than one million prescriptions), the
drug is assigned a yellow risk rating. This condition is based on
the understanding that the safety profile of a drug may be known
more completely only after being marketed for a length of time or
after a certain number of people have been prescribed the drug. The
values of two years and one million prescriptions can be modified
based on the knowledge gained through experience. For example, if
FDA were to increase the number of years a drug must be tested, or
the number of patients required in the phase III clinical studies,
prior to approval, those both could have an impact on the condition
set in step 325 because they both would be expected to provide more
safety profile information before approval. Obtaining more safety
information prior to approval and marketing is expected to change
the condition of step 325 by reducing the number of years on the
market and the number of prescriptions written. Of course, if
public pressure demands faster approval of drugs with less
regulatory oversight, it would be expected that both conditions of
step 325 would be increased.
[0063] It should be noted that in contrast to the assessments made
at steps 305, 310, and 315, the assessments made at steps 320 and
325 are not based on inputs from the algorithms 100 and 200 of
FIGS. 3 and 4, respectively. Instead, this is input derived from
other sources. For example, FDA designates certain drugs with the
orphan designation based on the size of the population that will be
prescribed the drug. The number of years of market experience can
be obtained from a number of sources, such as the FDA's web page
listing of approvals of drugs, which includes the date of approval.
The number of years of market experience and the number of
prescriptions written both can be obtained from information
providers such as IMS Health.
[0064] If the condition of step 325 is not met, then the algorithm
will determine the condition of step 330: does a threat exist with
a D threat level for the drug? As will be obvious by examining the
flow of algorithm 300, the algorithm will have already determined
whether a D threat level exists for a penetration value of both 85
and 1. Thus, step 330 effectively determines whether a D threat
level exists for the drug for a penetration value of 0 or greater.
If the condition of step 330 is met, the algorithm assigns a risk
rating of blue to the pharmaceutical.
[0065] If the condition of step 330 is not met, the algorithm then
determines the condition of step 335: does a threat with a C threat
level exist for the drug. The condition of step 335 is similar to
that of step 330 in that the algorithm has already determined
whether a C threat level exists for a penetration value of 85.
Thus, in the arrangement of algorithm 300, step 335 effectively
determines whether a C threat level exists for a penetration value
of 0 or greater. If a C threat level exists for a threat associated
with the drug, the algorithm assigns a risk rating of blue alert to
the pharmaceutical.
[0066] If the condition of step 335 is not met, the algorithm 300
then determines the condition of step 340: does a threat with a B
threat level exist for the drug. A B threat level indicates a lower
degree of risk than a C threat level, which in turn indicates a
lower level of risk than a D threat level. If a B threat level
exists for the drug, the algorithm assigns a blue risk rating to
the drug. If the condition of step 340 is not met, then the drug is
assigned a green alert.
[0067] It is believed by the inventor that applying the above
algorithms to pharmaceuticals and their associated threats will
provide users of pharmaceuticals with information that enables them
to make more informed decisions about taking their pharmaceutical.
In particular, the risk rating resulting from algorithm 300 of FIG.
5 is designed to provide patients the general risk rating and
guidance from Table IV. The guidance can be used as an action item
that patients can use when taking the pharmaceutical.
TABLE-US-00004 TABLE IV Relationship between Risk Rating, Risk
Advisory Conditions and Guidance to Patients Risk Advisory Risk
Rating Condition Guidance to the Patient RED HIGH Should only be
used under strict supervision ORANGE ELEVATED Use within a
customized risk management plan YELLOW GUARDED Be on the lookout
for safety events BLUE GENERAL Use under the normal care of a
physician GREEN LOW Suitable for widespread use
[0068] The algorithms described above can be modified to include
additional factors. In particular, the algorithms 10 and 300 of
FIGS. 1 and 5, respectively, can be modified to include additional
factors and variables. For example, the output of the algorithms 10
and/or 300 can be configured to include a step of assessing the
strength of the evidence used to support the safety alert.
Assessing the strength of the evidence can be based on using the US
Preventive Services Task Force (USPSTF) classification. The
strength of the evidence can be based, for example, on the
following four classifications: (i) evidence obtained from at least
one properly designed randomized controlled trial; (ii) evidence
obtained from well designed studies; (iii) opinions of respected
authorities; and (iv) dramatic implications of this alert, despite
lacking evidence. It is intended that other modifications of the
algorithms described herein can be made and remain within the scope
of the inventions.
[0069] The above algorithms are configured to process safety
information relating to a drug as applied to a total population of
patients administered the pharmaceutical. The algorithms also can
be configured to provide safety information for sub-populations
based on information provided about sub-populations and
individuals. The system can be used to provide personalized risk
ratings by recalculating penetration ratings for each threat given
the narrower population to which an individual belongs.
[0070] For example, a particular anti-hypertensive drug may be
taken by both men and women, of all ages, of all races, and in
people with and without diabetes. When applying the algorithm 100
to the general population this drug might be assigned a general
risk rating of blue. To assign a personalized risk rating to an
Asian male with diabetes aged 65 and older, one would re-assign a
penetration rating for each individual threat based not on the
general population but upon a population of Asian diabetic males
aged 65 and older. A particular threat that has a penetration
rating of 1 in the general population may have a penetration value
of 85 amongst elderly diabetic males. This may result in an
increased rating for the product for this narrower population, and
a personalized risk rating of orange for the Asian male with
diabetes aged 65 and older. As may be evident, only step 115 is not
dependent on the definition of the treatment population and the
result for the other steps in the algorithm 100 may vary depending
upon the definition of the population for whom the risk rating is
being determined.
[0071] The methods described herein may be implemented in a number
of manners. For example, the method of assigning a risk rating may
be implemented in software, on a computer system and/or on the
Internet. In one implementation, the system implementing the
algorithms may be stored and/or run on a central server which
gathers, processes, and stores threat information about medical
products. In such an implementation, the algorithms described
herein may be run on the central server. Alternatively, a user
accessing the central server may have software installed on a home
computer and the algorithms run on the home computer with data
supplied by the central server. Thus, if an individual desires to
know more about the safety and risk associated with the
pharmaceuticals they may go to the webpage of the system and enter
the name of one or more pharmaceuticals. The system then responds
back with the risk rating for each pharmaceutical.
[0072] The individual may further decide to obtain a more
personalized assessment of the risk rating associated with the
drug. The individual then would enter information about their
demographics (e.g., age, gender, race, weight, etc.) and their
health (e.g., hypertensive, elevated cholesterol levels, diabetic,
etc.). Depending upon the information already gathered for the
drug, the algorithms will be run as described above to assess a
more personalized risk rating for the individual.
[0073] In one implementation, the webpage and service described
above may allow the user to create a private account and specify
certain drugs for which they have been prescribed. Each time the
user accesses the account, the webpage may display the risk rating
for each of the drugs they have selected (e.g., an orange risk
rating for one drug, a blue risk rating for another drug, etc.)
along with a personalized risk rating for each of the drugs. In
this manner, individuals are able to obtain real-time overall and
personalized risk ratings for the pharmaceuticals they are being
prescribed.
[0074] In another implementation, physicians can use a
physician-oriented page of the web system to monitor the risk
rating of pharmaceuticals they prescribe to their patients along
with the personalized risk ratings for each of their patients. The
physician-oriented page can be configured to display in one screen
all the drugs prescribed by the physician as well as each drug's
risk rating. On another screen, the physician can access
pharmaceutical information for each patient and view personalized
risk ratings. This configuration will allow the physician to notify
their patients of potential risk as well as help the physician
prescribe pharmaceuticals with more knowledge about the risks and
potential threats associated with each drug and each patient.
EXAMPLE
[0075] The above methods and systems may be implemented in a
webpage as described above. In one example of such a webpage
implementation, the user logs onto an account and enters the name
for a particular drug. The risk rating displayed on the webpage for
the drug will be an indicator of the risk for the drug, e.g., a
yellow box adjacent to the drug name. The following example
describes one implementation for how a risk rating could be
assigned and the box displayed.
[0076] Consider a drug, NUSAFEX, used to prevent pregnancy and
which has been on the market for nine months. As a result of
clinical trials and spontaneous reports, the drug's label
identifies the following threats: (1) a generally increased
probability of death of approximately 1 in 250,000; (2) a small but
notably increased risk of bone fractures in young women; (3) a high
risk of death in patients with severe liver damage. Given its
branding and the many elements in its label, many patients may be
confused about how risky this drug is for them. Applying the
algorithms described in the figures above, one would initially
evaluate the threat levels using the algorithm of FIG. 4. Applying
the algorithm 200 of FIG. 4 to individually rate each of the
threats results in threat ratings of A, B, and D, respectively.
Applying the algorithm 100 of FIG. 3 to evaluate the penetration
rating for each of the individual threats, results in penetration
ratings of 85, 15, and 0, respectively. The first threat is general
to the entire treatment population, which is greater than 85% of
the treatment population and therefore is assigned a penetration
rating of 85. The second threat is specific to young women that
make up more than 15% of the treatment population (i.e., women
taking contraceptives) and therefore a penetration rating of 15 is
assigned. The third threat is specific to people with severe liver
damage that make up much less than 1% of the treatment population
and therefore is assigned a penetration rating of 0. This results
in individual threat ratings of A85, B15, and D0.
[0077] Applying the algorithm 300 of FIG. 5, the algorithm will
find a match at step 325 because NUSAFEX has been on the market for
less than two years and will result in a risk rating of yellow.
This allows a young woman considering contraceptive options to
balance the newer benefits of NUSAFEX against its yellow risk
rating to make a more informed decision about taking NUSAFEX
instead of an older contraceptive that lacks the new benefits but
has a blue rating. In the absence of the methods described herein,
the woman would not have the ability to make as informed a decision
without research.
[0078] As another example of how the algorithms described herein
can be used to generate a personalized risk rating, one could
imagine a nineteen year old female with chronic hepatitis and
significant liver dysfunction. Although the threat levels remain A,
B, and D, the penetration ratings for each individual threat in
this narrower treatment population of patients with chronic
hepatitis will now be 85.15, 85 since the third threat will apply
to more than 85% of patients in this treatment population. In
applying algorithm 300, the resulting risk rating for 19 year women
with chronic hepatitis and significant liver dysfunction will be
red because the newly calculated D85 will match at step 305. The 19
year old woman is able to easily comprehend the risk associated
with NUSAFEX without needing to look into details.
[0079] While several particular forms of the invention have been
illustrated and described, it will be apparent that various
modifications and combinations of the invention detailed in the
text and figures can be made without departing from the spirit and
scope of the invention. For example, references to utilities or
applications are also not intended to be limiting in any manner.
Similarly, references to specific indicators, colors, symbols,
letters, numbers and the like are exemplary only and may be varied
within the scope of the inventions. Accordingly, it is not intended
that the invention be limited, except as by the appended
claims.
* * * * *